Role of geophysics in glacial hazard assessment

Transcription

1 first break volume 24, August 2006 special topic Role of geophysics in glacial hazard assessment John M. Reynolds of Reynolds Geo-Sciences,* explains how geophysical methods can play an important role in understanding glacial hazards and their little discussed but potentially disastrous threat to communities in the world s mountainous regions and beyond. C limate change, whatever its causes, is resulting in the recession of glaciers (e.g. IPCC, 2001) and the development of glacial lakes in high mountain chains across the world. As time progresses, more lakes are forming and lake volumes are increasing. In parallel with this, catastrophic failure of terminal moraines damming such lakes is occurring increasingly frequently especially in the Himalayas, with recent examples also in the Alps and the North American Rockies, among other areas. In a typical glacial lake outburst flood (GLOF), some million m 3 of water and debris can be disgorged downstream causing widespread damage and destruction, in some cases for hundreds of kilometres. Peak discharges can exceed 10,000 m³/sec down river channels more often used to dealing with less than a few hundred m³/sec even during peak monsoon flows. Flood waves tens of metres high can travel at speeds of tens of kilometres per hour, with flood durations lasting from one to several days. The effects of such events can last for decades. What are glacial hazards? When a glacier or glacier-related feature adversely affects human activities, directly or indirectly, it is considered to be a hazard (Reynolds, 1992). This covers the range of human activities from hill farmers in Peru or Switzerland, for example, having their livelihoods (and lives) threatened by avalanches, through to the environmental effects of iceberg impacts on seabed oil installations off Labrador; from catastrophic glacier outbursts destroying Icelandic road communications, to glacier recessions resulting in increased storage of water behind fragile terminal moraine dams at high altitude such as in the Andes in Peru, the Himalayas, Tien Shan Mountains, Pamirs, etc. What is of growing concern in mountainous areas, especially in developing countries, is the Figure 1 Schematic diagram of a hazardous moraine-dammed glacial lake. Factors contributing to the hazards include (a) large lake volume, (b) narrow and high moraine dam, (c) stagnant glacier ice within the dam, and (d) limited freeboard between the lake level and the crest of the moraine ridge. Potential outburst flood triggers include avalanche displacement waves from (A) calving glaciers, (B) hanging glaciers, and (C) rockfalls; (D) settlement and/or piping within the dam (due to progressive seepage or seismic activity); (E) melting ice core; and (F) catastrophic drainage into the lake from sub-glacial or en-glacial channels or supra-glacial lakes (from Richardson and Reynolds, 2000). *Unit 17, Mold Business Park, Wrexham Road, Mold, Flintshire, CH7 1XP, UK; EAGE 61

2 special topic first break volume 24, August 2006 Figure 2 GLOF breach at Sabai Tsho, Nepal. The scale of the moraine and its collapse can be gauged by reference to the houses in the right foreground. range of hazards associated with pro- and supra-glacial lakes that can lead to devastating GLOFs. The components of a glacial lake system are shown in Figure 1 and these comprise two sets threshold features, such as the terminal and lateral moraines themselves and their composition and structure, and whether or not they contain stagnant melting glacier ice. The second set consists of features that might trigger a process that might lead to the failure of the moraine dam, such as by overtopping of the moraine by a wave arising from the displacement of lake water from a rock or ice avalanche. Such waves can be massive. For example, at Safuna Alta in the Cordillera Blanca, Peru, on 22 April 2002, a massive rock avalanche collapsed onto the local glacier and into its pro-glacial lake displacing virtually half the lake into a wave ultimately around 110 m high that overtopped the moraine dam and caused damage downstream (Hubbard et al., 2005). More usually, a much smaller displacement wave (2-5 m) can overtop a moraine leading to regressive erosion of the distal flank, creating a new outflow portal that through the high flow rates of water progressively enlarge the channel until such point that the moraine itself fails, which can be almost explosive. The resulting scars of moraine breaches can be most impressive (Figure 2). In other cases, melting stagnant glacier ice buried within the moraine causes subsidence of the dam ultimately leading to its physical failure. Historical events In the Himalayas, during the 1940s-1960s, one significant GLOF was reported to have occurred per decade. By the 1990s, it had become one event per three years; by 2010 it is expected that one causing more than $1 million dollars of damage will occur each year (Richardson and Reynolds, 2000). Should one of these GLOFs engulf a mature hydroelectric power facility that has been operating successfully for more than a decade, it is estimated that the cost of the damage could exceed $500 million. For a country such as Nepal, the fourth poorest in the world, such an event could set back its economic development by more than a decade, even threatening its very economic survival. In Peru, for example, a major hydroelectric power station that provides the majority of electricity to the city of Trujillo, the country s third largest city, is potentially vulnerable from GLOF damage (Reynolds, 1998). In 1998 the hydropower station at Macchupichu was destroyed by a glacially-induced debris flow, resulting in over $200 million in loss of revenue and rebuilding costs. Over 32,000 people have been killed in Peru alone over the last century due to these aluviones as they are known in South America (Reynolds, 1992) and many more are at risk. In 1994, Luggye Tsho in northern Bhutan ruptured releasing 18 million m³ that flowed over 200 km into northern India, killing 21 Bhutanese villagers at Punakha. A neighbouring glacial lake at Thorthormi Glacier is currently the most hazardous glacial lake in Bhutan and could release in excess of 50 million m³ along the same river channel, easily penetrating into India and potentially even into Bangladesh. Many glacial lakes in one country drain into rivers that flow into another country creating interesting geopolitical issues in managing the risk, but that is another story! The Himalayas are being seen as the new powerhouse of Central Asia, exporting energy into India and beyond. The development of such facilities also brings increased rural development and better infrastructure, drawing in people from the more rural areas as better communications make transport of goods to market easier. Consequently, these typically more than $1 billion scale projects result in concentrating infrastructure and communities into vulnerable valleys. Taking both the increasing occurrence of the glacial hazards with the growing vulnerability means that the overall risks in these environments is, in some cases, becoming critical, with thousands of lives and millions of Figure 3 3D perspective of Drogpa Nagtsang Glacier, Manlung Chhu, Tibet (China) derived from satellite images draped over a Digital Elevation Model. Such images can be used to aid the design of ground-based investigations EAGE

3 first break volume 24, August 2006 special topic Figure 4 Electrical resistivity tomogram at Imja Tsho, Nepal, showing the absence of high resistivity ice within the moraine. Zones of large boulders in the near surface are evident as high resistivity blobs in otherwise more conductive moraine. dollars worth of investments at risk. While much is being done, especially in Peru, to manage the risk, there are still many areas where the assessment of the initial glacial hazard is rudimentary at best. Hazard assessment Over the last decade or so, a great deal has been learnt both about how glacial hazards develop as well as methods by which to assess the hazards at various scales. Exciting new developments have been made using optical remote sensing (Kääb et al. 2003; Quincey et al. 2005) to facilitate the mapping of glaciers and glacial lakes over whole catchments and synthetic aperture radar remote sensing to measure rates of surface movement of debris-covered glaciers; potentially hazardous glacial lakes tend to develop on stagnating flat-lying parts of glaciers (Reynolds, 2000). This provides a method by which to identify those sites at which there should be ground investigations and where to target specific types of ground-based surveys, including geophysical investigations. For example, Figure 3 shows an extract of a pan-sharpened multi-spectral SPOT5 image draped over a digital elevation model for a developing glacial lake in Manlung Chhu in Tibet (China). Extracts such as this can be articulated on the computer to give various perspectives in order to help identify potential locations for geophysical and other transects to yield important information in the overall hazard assessment. In 2003, as a result of a three-year UK-government funded research project, the first guidelines for the assessment of glacial hazards, were published (RGSL, 2003). These include a section specifically on the use of geophysical methods in glacial hazards assessment (Appendix A4) and are available on the internet (PDFs can be downloaded from database/reports/colour/ ADD048_COL.pdf or from A comprehensive ground investigation would consist of a range of surveys, from basic topography, geomorphology, engineering geology, and glacial geology through to structural glaciological mapping of the remnant parent glacier. All of these effectively are used to map the ground surface and to deduce processes that may or may not impact onto the glacier and lake system. Interpretations from these are made to determine internal structures of the moraine dams with a view to considering the geotechnical stability of the moraine and its relationship with the lake or other water bodies. What is crucially missing in most cases is a measure of the third dimension, to image what is there rather than just hypothesise about it. This can make the difference between assessing the hazard correctly or not. Mistakes made at this stage can have huge implications for the lives of many as has happened tragically in the past. For this reason, geophysical surveys, when properly carried out and interpreted by people experienced in glacial geology, can play a critical role in the assessment of glacial hazards. The role of geophysics Geophysical methods have been used extensively over clean glaciers and ice sheets for many years. However, undertaking appropriate geophysical investigations over debris-covered glaciers and large moraines that might contain buried ice is an altogether different logistical and scientific challenge. Early attempts were made in the Alps and the Cordillera Blanca in Peru in the 1960s and 1970s (e.g. Röthlisberger, 1967; Lliboutry, 1977). In the Himalayas electrical resistivity tomography and a basic form of low-frequency radar were first used very effectively in 1996 at Thulagi, Manaslu District, Nepal (Delisle et al. 2003; Hanisch et al. 1998; 2006 EAGE 63

4 special topic first break volume 24, August 2006 Figure 5 Photo of GPR data being acquired over the ice apron in front of the actively calving Trakarding Glacier. The lake of Tsho Rolpa lies off to the right of this picture. Pant and Reynolds, 2000). This work led to the first use of a commercial radar system (Mala Geoscience RAMAC system) in Nepal at Tsho Rolpa in 1997 with further electrical resistivity tomography (Rana et al. 2000; Richardson and Reynolds, 2000). The geophysical work at the lateral and terminal moraines of Tsho Rolpa was considered to be critical to the overall assessment of the hazard and to the interim remediation strategy of the lake. The moraine damming the lake is about 150 m high; in 1994 there was a small natural outlet in the middle of the terminal moraine but elsewhere the freeboard was barely 1 m and the lake was literally lapping at the dam crest. Springs emanated from about one third of the way up on the distal moraine flank, and there was morphological evidence for the presence of buried stagnant ice within the northwest part of the moraine. Trakarding Glacier terminated at the eastern end of the lake some 3 km distant but was actively calving large blocks of ice producing small waves that reached the moraine dam. Several thousand people live in small villages downstream and Khimte II HEP is operational 86 km downstream. In 1992, it was clear that something had to be done to remediate this lake urgently. In 1995, experiments were undertaken using siphons, for the first time in the Himalayas. While these worked extremely effectively, they could not provide the capacity to draw down the lake level sufficiently within the time required. To cut a long story short, funding was found and work started in 1997 to design a remediation system that could take out sufficient water fast enough to provide a tangible reduction in risk. Part of this process was to answer the question as to where the buried ice was located within the moraine. To excavate into the core of the dam only to reveal buried ice was likely to trigger the collapse of the moraine dam and this was to be avoided at all costs. To this end, ground penetrating radar (GPR) was used extensively over the moraine to determine the physical structure of the dam to aid the geotechnical estimation of its stability, as well as to demonstrate that certain areas were clear of any buried ice. To a lesser extent, electrical resistivity tomography was also used and both methods, especially when used coincidently, provided very clear evidence for both the absence of ice as well as its presence elsewhere (Figure 4). Resistivity imaging was used also to test for possible piping through the moraine linking with seepages at a lower level. Another consideration was the status of the glacier tongue feeding into the lake was it afloat or was it grounded? If it was the former, to reduce the lake level suddenly could have had disastrous consequences should the ice tongue break off great blocks of many thousands of tonnes the ensuing displacement wave would have destroyed any remediation work and would have breached the dam. So GPR was used on an exposed ice apron at the base of the ice cliff of the glacier tongue to try to answer this question (Figure 5). Further evidence was also sought by additional surface geological mapping on and around the glacier itself. This, coupled with a detailed structural glaciological analysis of the glacier, enabled the production of a model for Figure 6 Completed open cut channel and sluice gates at Tsho Rolpa, Nepal, after the lake level had been drawn down successfully by 3.5 m in July The portion of the moraine at the upper right hand side of the picture is cored with stagnant glacier ice EAGE

5 first break volume 24, August 2006 special topic (a) (b) Figure 7 (a) Imja Tsho and its debris-covered stagnant ice area and terminal and lateral moraines, with (b) GPR and (c) electrical resistivity results across the profile line shown in red in (a). Areas of very high resistivity indicate where buried ice is present. (c) the behaviour of the glacier tongue under ambient conditions and also for when the lake draw down had commenced. As a direct consequence of the use of geophysical methods at the terminal moraine, what was considered to be the most suitable location for an open cut channel was identified. Work began in 1999 and the channel and sluice gate were installed successfully and the lake level drawn down without incident by 3.5 m in July 2000 (Figure 6). The GPR and resistivity data revealed buried ice 20 m thick in the northwest part of the terminal moraine and active thermokarst degradation and subsidence is ongoing. In 1997, in part as a consequence of the geophysical results, it was recommended that the lake level should be drawn down a total of 20 m; the first 3.5 m has been accomplished but funds to achieve the remaining work have still to be found. Meanwhile the buried ice in the dam continues to melt Detailed geophysical investigations have also been undertaken at Imja Tsho, near Mt Everest in Nepal. These were carried out both to assist with the assessment of the hazard at this lake but also to test the efficacy of these methods (Figure 7). The results indicated that this lake, while large, is currently well constrained by its moraine and is not regarded as posing a serious hazard in the short term. Other geophysical methods have been deployed at other glacial lakes (seismic reflection, refraction, micro-gravity, and electro-magnetic profiling). None of these has produced sufficiently good quality information to justify the expense and physical effort of undertaking them. There are huge difficulties in placing geophones over boulder-rich moraine and in obtaining sufficient ground coupling of the energy source to achieve anything like reasonable depth penetration. The topographic variability across hummocky moraines (up to 2006 EAGE 65

6 special topic first break volume 24, August 2006 ±20 m is not unusual) and the heterogeneity of the materials makes micro-gravity too slow to acquire and too difficult to obtain acceptable corrections to the observed gravity data to produce reasonable Bouguer Anomaly data that can be modelled realistically. EM surveys are also too badly affected by the terrain and also lack the resolution to be useful in what is a very low conductivity environment. It is clear that the best and most practical methods across debris-covered glaciers and moraines are GPR and electrical resistivity tomography. The self potential method has also been demonstrated to be useful in mapping seepages through natural dams. Conclusions Geophysical methods, especially GPR and electrical resistivity tomography, have been demonstrated to be highly effective in mapping especially the presence of buried ice within terminal moraines as well as possible areas of piping through the dams. Being able to obtain vital 3D information is in cases critical to the best estimation of glacial hazards. To this end geophysical techniques have come of age and are now recommended as an integral part of the ground-based exploration and mapping suite. Not only is this information useful for the assessment of hazards but has also been proven in providing information essential to the design of safe and effective remedial works. As a consequence of the latest methods in glacial hazard assessment and appropriate remediation, there are many thousands of people whose safety has been improved immeasurably and vital infrastructure installed in ways in which risks have been reduced significantly. The role of appropriate geophysical methods has now been well established in the assessment of glacial hazards and in the subsequent design of remediation. Acknowledgements The author acknowledges the valuable contributions to this work by Dr Shaun Richardson, formerly RGSL, Dr Surendra Pant, Tribhuvan University, Nepal, and colleagues at the Department of Hydrology and Meteorology, His Majesty s Government of Nepal, Kathmandu. Thanks are also due to colleagues at the Centre for Glaciology, University of Wales, Aberystwyth, and Glaciología y Hydrología, INRENA, Huaraz, Peru, for ongoing collaboration on glacial hazard assessment and remote sensing. References Delisle, G., Reynolds, J.M., Hanisch, J., Pokhrel, A.P., and Pant, S. [2003] Lake Thulagi/Nepal: rapid landscape evolution in reaction to climate change. (Z. Geomorph. N.F.) Annals of Geomorphology, 130, 1-9. Hanisch, J., Delisle, G., Pokhrel, A.P., Dixit, A.M., Reynolds, J.M., and Grabs, W.E. [1998] The Thulagi glacier lake, Manaslu Himal, Nepal - Hazard assessment of a potential outburst. Proceedings of the Eighth International IAEG Congress, September 1998, Vancouver, Canada, Hubbard, B, Heald, A., Reynolds, J.M., Quincey, D., Richardson, S.D., Zapata, M., Santillan, N., and Hambrey, M.J. [2005] Impact of a rock avalanche on a moraine-dammed proglacial lake: Laguna Safuna Alta, Cordillera Blanca, Peru. Earth Surface Processes and Landforms, 30, Intergovernmental Panel on Climate Change (IPCC). [2001] Third assessment report. Working Group I. Cambridge University Press, Cambridge, UK. Lliboutry, L. [1977] Glaciological problems set by the control of dangerous lakes in Cordillera Blanca, Peru. II. Movement of covered glacier embedded within a rock glacier. Journal of Glaciology, 18, 79, Kääb, A., Wessels, R.L., Haeberli, W., Huggel, C., Kargel, J.S., and Khalsa, S.J.S. [2003] Rapid ASTER imaging facilitates timely assessment of glacier hazards and disasters. EOS, Transactions of the American Geophysical Union, 84, Pant, S.R. and Reynolds, J.M. [2000] Application of electrical imaging techniques for the investigation of natural dams: an example from the Thulagi Glacier Lake, Nepal. Journal of the Nepal Geological Society, 22, Quincey, D.J., Lucas, R.M., Richardson, S.D., Glasser, N.F., Hambrey, M.J., and Reynolds, J.M. [2005] Optical remote sensing techniques in high-mountain environments: application to glacial hazards. Progress in Physical Geography, 29, 4, Rana, B., Shrestha, A.B., Reynolds, J.M., Aryal, R., Pokhrel, A.P., and Budhathoki, K.P. [2000] Hazard assessment of the Tsho Rolpa Glacier Lake and ongoing remediation measures. Journal of the Nepal Geological Society, 22, Reynolds Geo-Sciences Ltd (RGSL). [2003] Development of glacial hazard and risk management protocols in rural environments Guidelines for the management of glacial hazards and risks. RGSL, Mold, UK. Reynolds, J.M. [1992] The identification and mitigation of glacier-related hazards: examples from the Cordillera Blanca, Peru. In: McCall, G.J.H., Laming, D.C.J. and Scott, S. (Eds), Geohazards, London, Chapman & Hall, Reynolds, J.M. [1998] Managing the risks of glacial flooding at hydro plants. Hydro Review Worldwide, 6, 2, Reynolds, J.M. [2000] On the formation of supraglacial lakes on debris-covered glaciers. In: Nakawo, M., Raymond, C.F. and Fountain, A. (Eds.), Debris-Covered Glaciers, Proceedings of the Seattle Workshop, USA, September 2000, IAHS Publication 264, Richardson, S.D. and Reynolds, J.M. [2000] An overview of glacial hazards in the Himalayas. Quaternary International, 65/66, Röthlisberger, H. [1967] Electrical resistivity measurements and soundings on glaciers: introductory remarks. Journal of Glaciology, 6, 47, EAGE